I read with interest the Maverick Citizen article by Vardas and colleagues, “Antibody testing – the missing weapon in the fight against Covid-19 in South Africa”, which provided excellent background into the two types of tests that could be used in identifying someone infected by the coronavirus SARS-CoV-2, and some of the history behind South Africa’s testing strategy. 

Painting antibody testing as a “missing weapon” in the fight against Covid-19 is, however, misleading and fails to send the clear messages the current scientific data is providing us with in relation to when antibody tests may be helpful and, more important, when they are not.

Simply put, tests used in Covid-19 either identify the virus in the airways employing the polymerase chain reaction (PCR) to identify SARS-CoV-2 genes, or an immune response mounted by the infected person as their immune system responds to the virus. The marker of the immune response we use is the production of proteins called antibodies that can be found in blood. These antibodies recognise and hopefully neutralise the virus in the body. Serology is a term used for tests that detect these antibodies.

The polymerase chain reaction is positive in the days immediately leading up to the appearance of symptoms and for about the first week of the illness, during which time the amount of virus detectable in the airways falls off. Therefore, the polymerase chain reaction test is helpful for doctors to confirm their clinical suspicion that the person in front of them manifesting symptoms suggestive of Covid-19 has the infection. As the polymerase chain reaction becomes negative early in the symptomatic period in the majority of people, the polymerase chain reaction cannot be used for telling us if someone has had a past infection with SARS-CoV-2.

At the present time we have very little data to guide us on how reliable current serology tests are in detecting antibodies after 35 days from symptom onset, which may impact on surveillance studies, and more important, what level of protection against reinfection the presence of antibodies may predict. Further research is needed in both these areas.

In contrast to the usefulness of the polymerase chain reaction test to diagnose acute Covid-19, our immune response takes longer to kick in once we’ve been infected by the virus. Antibodies are slow to develop and be detected in blood; international and local experience with serology tests for SARS-CoV-2 has confirmed that antibodies are not reliably picked up by currently available tests in the first two weeks of symptoms.

The chance of a positive result when a person with Covid-19 presents with symptoms is of less benefit than tossing a coin i.e. it’s about as likely to deny that the person had Covid-19 as it is to confirm it. For this reason, I strongly disagree with the opinion of the authors that “serological tests for SARS-CoV-2 can play an important role in the diagnosis and management of Covid-19”. Serology is not a test that helps doctors diagnose or treat people with Covid-19, with the exception of a tiny number of children who have been found to have a hyperinflammatory syndrome related to SARS-CoV-2, where the polymerase chain reaction is often negative, but serology positive. But the numbers of such cases are minuscule in the grand scheme of Covid-19, and a specific exception.

So, is there a place for serology testing in South Africa’s response to the epidemic? The answer is a likely yes, but only in the surveillance arena, enabling large-scale mapping of infections to show the rate of previous infection with the virus, and in the research arena.

At the present time we have very little data to guide us on how reliable current serology tests are in detecting antibodies after 35 days from symptom onset, which may impact on surveillance studies, and more important, what level of protection against reinfection the presence of antibodies may predict. Further research is needed in both these areas.

Rather than being the “missing link” in South Africa’s Covid-19 response, serology testing at the current time plays very little role in mitigating South Africa’s Covid-19 epidemic and on current evidence, is certainly not to be recommended as a diagnostic test for acute Covid-19. DM/MC

Marc Mendelson is Professor of Infectious Diseases at the University of Cape Town.